Balazs Harangi

Automatic detection of the optic disc using majority voting in a collection of optic disc detectors

This paper proposes an efficient method for locating the optic disc in retinal images automatically using majority voting scheme and data fusion. We show that instead of inventing a new algorithm which ends up being a minor variation on an old idea, the fusion of different optic disc (OD) detectors can enhance the overall performance of the detection system. The optic disc centre candidates of different optic disc detectors are marked in the image and a circular template is fit on each pixel in the image to count the outputs of these algorithms that fall within the radius. The location with maximum number of optic disc centre candidates is the hotspot and is used to localize the optic disc centre. An assessment of the performance of the combined optic disc detector versus detectors working separately is also presented. Our method achieved highest performance (overall 100% correct detection).

Automatic exudate detection using active contour model and regionwise classification

Diabetic retinopathy is one the most common cause of blindness in the world. Exudates are among the early signs of this disease, so its proper detection is a very important task to prevent consequent effects. In this paper, we propose a novel approach for exudate detection. First, we identify possible regions containing exudates using grayscale morphology. Then, we apply an active contour based method to minimize the Chan-Vese energy to extract accurate borders of the candidates. To remove those false candidates that have sufficient strong borders to pass the active contour method we use a regionwise classifier. Hence, we extract several shape features for each candidate and let a boosted Naïve Bayes classifier eliminate the false candidates. We considered the publicly available DiaretDB1 color fundus image set for testing, where the proposed method outperformed several state-of-the-art exudate detectors.

Graph based detection of optic disc and fovea in retinal images

Diabetic retinopathy (DR) is the damage to the eyes retina that occurs with long-term diabetes, which can eventually lead to blindness. Screening programs for DR are being introduced, however, an important prerequisite for automation is the accurate localization of the main anatomical features in the image, notably the optic disc (OD) and the macula. A series of interesting algorithms have been proposed in the recent past and the performance is generally good, but each method has situations, where it fails. This paper presents a novel framework for automatic detection of optic disc and macula in retinal fundus images using a combination of different optic disc and macula detectors represented by a weighted complete graph. A node pruning procedure removes the worst vertices of the graph by satisfying predefined geometric constraints and get best possible detector outputs to be combined using a weighted average. The extensive tests have shown that combining the predictions of multiple detectors is more accurate than any of the individual detectors making up the ensemble.

Automatic exudate detection by fusing multiple active contours and regionwise classification

In this paper, we propose a method for the automatic detection of exudates in digital fundus images. Our approach can be divided into three stages: candidate extraction, precise contour segmentation and the labeling of candidates as true or false exudates. For candidate detection, we borrow a grayscale morphology-based method to identify possible regions containing these bright lesions. Then, to extract the precise boundary of the candidates, we introduce a complex active contour-based method. Namely, to increase the accuracy of segmentation, we extract additional possible contours by taking advantage of the diverse behavior of different pre-processing methods. After selecting an appropriate combination of the extracted contours, a region-wise classifier is applied to remove the false exudate candidates. For this task, we consider several region-based features, and extract an appropriate feature subset to train a Naïve-Bayes classifier optimized further by an adaptive boosting technique. Regarding experimental studies, the method was tested on publicly available databases both to measure the accuracy of the segmentation of exudate regions and to recognize their presence at image-level. In a proper quantitative evaluation on publicly available datasets the proposed approach outperformed several state-of-the-art exudate detector algorithms.

Automatic exudate detection with improved Naïve-bayes classifier

Nowadays diabetic retinopathy is one of the most common reasons of blindness in the world. Exudates are the primary sign of this disease so the proper detection of these lesions is an essential task in an automatic screening system. In this paper, we propose a method for exudate detection which performs with high accuracy. First, we identify possible regions containing exudates using grayscale morphology. Then, we extract more than 50 descriptors for each candidate pixel to classify them. We analyzed the information content of the descriptors and selected the most relevant ones. The selected features are used to train a boosted naïve Bayes classifier. We tested this approach on the publicly available DiaretDB color fundus image database, where the proposed detector outperformed the state-of-the-art ones regarding the FScore.

Detection of the optic disc in fundus images by combining probability models

In this paper, we propose a combination method for the automatic detection of the optic disc (OD) in fundus images based on ensembles of individual algorithms. We have studied and adapted some of the state-of-the-art OD detectors and finally organized them into a complex framework in order to maximize the accuracy of the localization of the OD. The detection of the OD can be considered as a single-object detection problem. This object can be localized with high accuracy by several algorithms extracting single candidates for the center of the OD and the final location can be defined using a single majority voting rule. To include more information to support the final decision, we can use member algorithms providing more candidates which can be ranked based on the confidence ordered by the algorithms. In this case, a spatial weighted graph is defined where the candidates are considered as its nodes, and the final OD position is determined in terms of finding a maximum-weighted clique. Now, we examine how to apply in our ensemble-based framework all the accessible information supplied by the member algorithms by making them return confidence values for each image pixel. These confidence values inform us about the probability that a given pixel is the center point of the object. We apply axiomatic and Bayesian approaches, as in the case of aggregation of judgments of experts in decision and risk analysis, to combine these confidence values. According to our experimental study, the accuracy of the localization of OD increases further. Besides single localization, this approach can be adapted for the precise detection of the boundary of the OD. Comparative experimental results are also given for several publicly available datasets.

Improving the accuracy of optic disc detection by finding maximal weighted clique of multiple candidates of individual detectors

In this paper, we propose a method to improve the automatic detection of the optic disc on fundus images. We have studied and implemented some of the optic disc detectors from concerning literature to organize them into an ensemble system. As a former work, we proposed an ensemble-based optic disc detection system, based on simple majority voting which already outperformed the individual detectors. To improve further the performance of the ensemble-based system, now we examine how we can extract more candidates from the individual algorithms to have the appropriate location of the optic disc among them. We also assign weights to each candidate based on the priority suggested by the algorithms. We consider these weighted candidates as vertices of a graph and look for a subgraph with a maximal sum of weights constrained by the geometry of the optic disc. Experimental results are also presented to see the improvement.

Detection of exudates in fundus images using a Markovian segmentation model

Diabetic retinopathy (DR) is one of the most common causing of vision loss in developed countries. In early stage of DR, some signs like exudates appear in the retinal images. An automatic screening system must be capable to detect these signs properly so that the treatment of the patients may begin in time. The appearance of exudates shows a rich variety regarding their shape and size making automatic detection more challenging. We propose a way for the automatic segmentation of exudates consisting of a candidate extraction step followed by exact contour detection and region-wise classification. More specifically, we extract possible exudate candidates using grayscale morphology and their proper shape is determined by a Markovian segmentation model considering edge information. Finally, we label the candidates as true or false ones by an optimally adjusted SVM classifier. For testing purposes, we considered the publicly available database DiaretDB1, where the proposed method outperformed several state-of-the-art exudate detectors.

Improving automatic exudate detection based on the fusion of the results of multiple active contours

In this paper, we recommend an algorithm for exudate detection which is based on the combination of active contours obtained for the image after different preprocessing algorithms. First, we extract the candidate regions of exudates with grayscale morphology and then we minimize the Chan-Vese energy function to achieve the nearly accomplished boundary on nine different pre-processed images. Considering the combinations of the nine regions derived from the results of the active contours we select the one as the final exudate region that meets the most of some textural constraints. For this task, we apply machine learning based classification with using a training data set to determine these constraints. We considered the publicly available DiaretDB1 color fundus image set, where our method achieved higher accuracy than other state-of-the-art exudate detectors regarding both sensitivity and correct boundary extraction.

Measuring regularity of network patterns by grid approximations using the LLL algorithm

In a recent work, we have proposed a novel way to approximate point sets with grids using the LLL algorithm, which operates in polynomial time. Now, we show how this approach can be applied to pattern recognition purposes with interpreting the rate of approximation as a new feature for regularity measurement. Our practical problem is the characterization of pigment networks in skin lesions. For this task we also introduce a novel image processing method for the extraction of the pigment network. Then, we show how our grid approximation framework can be applied with specializing it for the recognition of hexagonal patterns. The classification performance of our approach for the pigment network characterization problem is measured on a database annotated by a clinical expert. Throughout the paper we address several practical issues that may help to apply our general framework to other practical tasks, as well.